Check patentability & draft patents in minutes with Patsnap Eureka AI!

Image data-based pedestrian detection method

A technology of pedestrian detection and image data, applied in the field of pedestrian detection, can solve problems such as slow methods, and achieve the effect of improving the detection rate and speeding up the detection speed

Active Publication Date: 2017-11-07
HARBIN INST OF TECH SHENZHEN GRADUATE SCHOOL
View PDF4 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The Selective Search method used by RCNN and DeepPed to detect pedestrians extracts thousands of candidate boxes in a picture, and each candidate box must be sent to the neural network for judgment, so this method is very slow

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Image data-based pedestrian detection method
  • Image data-based pedestrian detection method
  • Image data-based pedestrian detection method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0014] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0015] like figure 1 As shown, a pedestrian detection method based on image data mainly combines the traditional HOG (Histogram of Oriented Gradient) + SVM (Support Vector Machine, Support Vector Machine) pedestrian detection method and low-rank sparse matrix decomposition method. Among them, the low-rank sparse matrix factorization adopts the GoDec model to solve this problem. The purpose of the GoDec model is that for a matrix, it can be processed by an algorithm and divided into three parts: the low-rank part, the discrete part and the noise:

[0016] The main modification of the present invention to Fast YOLO has three parts: data preprocessing, modification of Fast YOLO network structure and realization of the final Loss Function.

[0017] First, the network structure of GoogLeNet is shown in Table 4-1.

[0018] Table 4-1 GoogLeNet struc...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The present invention provides an image data-based pedestrian detection method. The image data-based pedestrian detection method comprises the following steps of (1), for the pre-processing of the data, adding a layer in the first layer of a network structure to read the data; (2) using the convolutional layers to substitute for an original YOLO full-connecting layer, adding an ReLU layer behind each convolutional layer, and then using a Reshape layer to change the inputted dimensions; (3) realizing a Loss Function. The beneficial effects of the present invention are that a Fast YOLO is realized through a Caffe interface and GoogLeNet-based network, and the full-connecting layer in an original network is changed into the convolutional layers, so that the detection speed is accelerated; moreover, the experimental results show that the detection rate also can be improved.

Description

technical field [0001] The invention relates to a pedestrian detection method, in particular to a pedestrian detection method based on image data. Background technique [0002] Pedestrian detection based on HOG and SVM is a very classic detection model, and HOG has been proven to be a very effective descriptor for human detection. [0003] Many of today's popular object detections are based on neural networks, which are not sensitive to window size. When RCNN and DeepPed detect pedestrians, they use a method called SelectiveSearch to perform some processing on the input image, and then send the processed results to the neural network. These image detection networks have a relatively high status in the field of face recognition. [0004] Pedestrian detection methods based on HOG and SVM are sensitive to the size of the detection window. Sometimes the videos shot based on low-altitude flying platforms have a large gap in the size of pedestrians, so this method is not applica...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06K9/00G06N3/08
CPCG06N3/084G06V20/40
Inventor 叶允明李旭涛李彦良夏武
Owner HARBIN INST OF TECH SHENZHEN GRADUATE SCHOOL
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com